A Heterocyclic Polyurethane with Enhanced Self-Healing Efficiency and Outstanding Recovery of Mechanical Properties

A functional polyurethane based on the heterocyclic group was synthesized and its self-healing and mechanical properties were examined. To synthesize a heterocyclic polyurethane, a polyol and a heterocyclic compound with di-hydroxyl groups at both ends were blended and the blended solution was reacted with a crosslinker containing multiple isocyanate groups. The heterocyclic polyurethane demonstrates better self-healing efficiency than the conventional polyurethane with no heterocyclic groups. Furthermore, unlike the conventional self-healing materials, the heterocyclic polyurethane examined in this study shows an outstanding recovery of the mechanical properties after the self-healing process. These results are attributed to the unique supramolecular network resulting from the strong hydrogen bonding interaction between the urethane group and the heterocyclic group in the heterocyclic polyurethane matrix.


Introduction
As display technology has evolved from flat displays to flexible (foldable, rollable, wearable, or stretchable) displays [1][2][3][4][5][6], the demand for new functional materials for flexible displays has increased significantly, particularly for surface protection. While hard coating materials have been extensively used to protect a flat display surface from mechanical surface damage such as dents or scratches (linear tearing of the surface), the properties of these coatings prevent their adaptation for use in flexible displays. For example, during folding, the inside surface of the folded display will be compressed while the outside surface will be extended. Because a hard coating layer is usually located at the outermost layer of the display, repeated folding and unfolding will generate significant stress accumulation on the hard coating layer. As conventional hard coating materials are known to be fragile and easily broken by mechanical stress, this cumulative stress will cause severe mechanical damage to the display surface and therefore will decrease the life expectancy of the displays. Self-healing polymeric materials are an emerging class of smart functional materials and have been widely studied during the past

Characterization
The optical properties of the samples were measured using a spectrophotometer (CM-3600d, Konica Minolta, Tokyo, Japan). The thermal properties were analyzed by differential scanning calorimetry (DSC) (DSC Q20, TA Instrument, New Castle, DE, USA) with a heating rate of 10 • C/min under nitrogen atmosphere. The structural analysis was performed using 1 H nuclear magnetic resonance (NMR) spectroscopy (VNMRS 600 MHz, Varian, Crawley, UK) with CDCl 3 as the NMR solvent, gel permeation chromatography (GPC) (Waters GPC system, Waters, Milford, MA, USA) equipped with a refractive index detector using tetrahydrofuran as the eluent with the columns calibrated against the standard polystyrene samples, and Fourier-transform infrared spectroscopy (FT-IR) (Nicolet iS50, Thermo Fisher Scientific, Waltham, MA, USA) using a spectrometer equipped with attenuated total reflectance accessory. The self-healing properties were measured using a nanoscratch tester (NST 3 , Anton paar, Graz, Austria) equipped with a conical tip with the size of 20 µm and a white light interferometer (NewView TM , ZYGO, Berwyn, PA, USA) at a loading speed of 2 mm/min. The mechanical properties were measured using a universal tensile machine (QM100SE, QMESYS, Gyeonggi, Korea) at a rate of 10 mm/min and a cell load of 10 kN. All of the specimens were prepared according to ASTM D1708: dog-bone in shape; 38 mm × 15 mm overall in size; 5 mm × 22 mm in the gauge area. Table 1 shows the designations and chemical structures of the samples in this study (see Figure  S1 in the Supplementary Information for the overall synthetic route, 1 H nuclear magnetic resonance (NMR) spectrum, and gel permeation chromatography (GPC) results for polyol). As a reference, conventional polyurethane (PU) was prepared by reacting a polyol solution and a crosslinker solution. A series of self-healable polyurethanes (PUB, PUT, and PUD) were prepared by reacting a polyol solution, crosslinker solution, and functional diol solution where 1,4-benzenedimethanol (BHMB), 2,5-bishydroxymethyl tetrahydrofuran (BHMTHF), and 1,2-dihydroxyethane (DHE) were used as the aromatic functional diol, heterocyclic functional diol, and aliphatic functional diol, respectively. It is noted that all of the functional diol solutions have identical OH values, and the theoretical degree of crosslinking of all of the samples was set to 100%. The thermal and optical properties were analyzed using a differential scanning calorimeter (DSC), a thermal gravimetric analyzer (TGA), and a spectrophotometer (see Figure S2-S4 in the Supplementary Information for the DSC thermograms and TGA curves and the detailed optical properties of PU, PUB, PUT, and PUD). Comparison of PU to PUB, PUT, and PUD shows that all of the samples have similar transition temperatures and also show high transmittance and low haze and yellow index values, indicating that all of the functional diols used in this study did not significantly affect the initial thermal and optical properties of PU as listed in Table 2. It is generally known that a residual solvent in polymeric system can act as a plasticizer and significantly affect the final glass transition temperature. When the TGA results were analyzed to measure the amount of residual solvent, all of the TGA curves showed no observable change in weight, which indicates that all of the samples used in this study have no residual solvents. properties were analyzed using a differential scanning calorimeter (DSC), a thermal gravimetric analyzer (TGA), and a spectrophotometer (see Figure S2-S4 in the Supplementary Information for the DSC thermograms and TGA curves and the detailed optical properties of PU, PUB, PUT, and PUD). Comparison of PU to PUB, PUT, and PUD shows that all of the samples have similar transition temperatures and also show high transmittance and low haze and yellow index values, indicating that all of the functional diols used in this study did not significantly affect the initial thermal and optical properties of PU as listed in Table 2. It is generally known that a residual solvent in polymeric system can act as a plasticizer and significantly affect the final glass transition temperature. When the TGA results were analyzed to measure the amount of residual solvent, all of the TGA curves showed no observable change in weight, which indicates that all of the samples used in this study have no residual solvents.  Table 2. Optical properties (transmittance (Tr), yellow index (YI), and haze), thermal property (glass transition temperature (Tg)), and thicknesses of conventional polyurethane (PU), PU with aromatic moiety (PUB), PU with heterocyclic moiety (PUT), and PU with aliphatic moiety (PUD). All of the data points were obtained by averaging the results of three measurements. A self-healing experiment was conducted for PU, PUB, PUT, and PUD using the single scratch technique. Figure 1 shows the optical microscopy images before and after the self-healing process. It should be noted that PUB, PUT, and PUD contain an aromatic benzene moiety, heterocyclic tetrahydrofuran moiety, and aliphatic ethylene moiety, respectively, where all of these moieties have approximately the same molecular size. Interestingly, while PU, PUB, and PUD still retained traces of damage even after the self-healing process, PUT successfully self-healed the damaged area and recovered to its initial state within a short time. To quantitatively evaluate the self-healing efficiency of PU and PUT, a nano/micro scratching technique was used. As shown in Figure 2, PUT shows perfect self-healing performance (0.1-0 μm, 100%) in contrast to PU (0.11-0.4 μm, 64%), indicating properties were analyzed using a differential scanning calorimeter (DSC), a thermal gravimetric analyzer (TGA), and a spectrophotometer (see Figure S2-S4 in the Supplementary Information for the DSC thermograms and TGA curves and the detailed optical properties of PU, PUB, PUT, and PUD). Comparison of PU to PUB, PUT, and PUD shows that all of the samples have similar transition temperatures and also show high transmittance and low haze and yellow index values, indicating that all of the functional diols used in this study did not significantly affect the initial thermal and optical properties of PU as listed in Table 2. It is generally known that a residual solvent in polymeric system can act as a plasticizer and significantly affect the final glass transition temperature. When the TGA results were analyzed to measure the amount of residual solvent, all of the TGA curves showed no observable change in weight, which indicates that all of the samples used in this study have no residual solvents.  Table 2. Optical properties (transmittance (Tr), yellow index (YI), and haze), thermal property (glass transition temperature (Tg)), and thicknesses of conventional polyurethane (PU), PU with aromatic moiety (PUB), PU with heterocyclic moiety (PUT), and PU with aliphatic moiety (PUD). All of the data points were obtained by averaging the results of three measurements.

Optical Properties
Thermal A self-healing experiment was conducted for PU, PUB, PUT, and PUD using the single scratch technique. Figure 1 shows the optical microscopy images before and after the self-healing process. It should be noted that PUB, PUT, and PUD contain an aromatic benzene moiety, heterocyclic tetrahydrofuran moiety, and aliphatic ethylene moiety, respectively, where all of these moieties have approximately the same molecular size. Interestingly, while PU, PUB, and PUD still retained traces of damage even after the self-healing process, PUT successfully self-healed the damaged area and recovered to its initial state within a short time. To quantitatively evaluate the self-healing efficiency of PU and PUT, a nano/micro scratching technique was used. As shown in Figure 2, PUT shows perfect self-healing performance (0.1-0 μm, 100%) in contrast to PU (0.11-0.4 μm, 64%), indicating properties were analyzed using a differential scanning calorimeter (DSC), a thermal gravimetric analyzer (TGA), and a spectrophotometer (see Figure S2-S4 in the Supplementary Information for the DSC thermograms and TGA curves and the detailed optical properties of PU, PUB, PUT, and PUD). Comparison of PU to PUB, PUT, and PUD shows that all of the samples have similar transition temperatures and also show high transmittance and low haze and yellow index values, indicating that all of the functional diols used in this study did not significantly affect the initial thermal and optical properties of PU as listed in Table 2. It is generally known that a residual solvent in polymeric system can act as a plasticizer and significantly affect the final glass transition temperature. When the TGA results were analyzed to measure the amount of residual solvent, all of the TGA curves showed no observable change in weight, which indicates that all of the samples used in this study have no residual solvents.  Table 2. Optical properties (transmittance (Tr), yellow index (YI), and haze), thermal property (glass transition temperature (Tg)), and thicknesses of conventional polyurethane (PU), PU with aromatic moiety (PUB), PU with heterocyclic moiety (PUT), and PU with aliphatic moiety (PUD). All of the data points were obtained by averaging the results of three measurements. A self-healing experiment was conducted for PU, PUB, PUT, and PUD using the single scratch technique. Figure 1 shows the optical microscopy images before and after the self-healing process. It should be noted that PUB, PUT, and PUD contain an aromatic benzene moiety, heterocyclic tetrahydrofuran moiety, and aliphatic ethylene moiety, respectively, where all of these moieties have approximately the same molecular size. Interestingly, while PU, PUB, and PUD still retained traces of damage even after the self-healing process, PUT successfully self-healed the damaged area and recovered to its initial state within a short time. To quantitatively evaluate the self-healing efficiency of PU and PUT, a nano/micro scratching technique was used. As shown in Figure 2, PUT shows perfect self-healing performance (0.1-0 μm, 100%) in contrast to PU (0.11-0.4 μm, 64%), indicating properties were analyzed using a differential scanning calorimeter (DSC), a thermal gravimetric analyzer (TGA), and a spectrophotometer (see Figure S2-S4 in the Supplementary Information for the DSC thermograms and TGA curves and the detailed optical properties of PU, PUB, PUT, and PUD). Comparison of PU to PUB, PUT, and PUD shows that all of the samples have similar transition temperatures and also show high transmittance and low haze and yellow index values, indicating that all of the functional diols used in this study did not significantly affect the initial thermal and optical properties of PU as listed in Table 2. It is generally known that a residual solvent in polymeric system can act as a plasticizer and significantly affect the final glass transition temperature. When the TGA results were analyzed to measure the amount of residual solvent, all of the TGA curves showed no observable change in weight, which indicates that all of the samples used in this study have no residual solvents.  Table 2. Optical properties (transmittance (Tr), yellow index (YI), and haze), thermal property (glass transition temperature (Tg)), and thicknesses of conventional polyurethane (PU), PU with aromatic moiety (PUB), PU with heterocyclic moiety (PUT), and PU with aliphatic moiety (PUD). All of the data points were obtained by averaging the results of three measurements. A self-healing experiment was conducted for PU, PUB, PUT, and PUD using the single scratch technique. Figure 1 shows the optical microscopy images before and after the self-healing process. It should be noted that PUB, PUT, and PUD contain an aromatic benzene moiety, heterocyclic tetrahydrofuran moiety, and aliphatic ethylene moiety, respectively, where all of these moieties have approximately the same molecular size. Interestingly, while PU, PUB, and PUD still retained traces of damage even after the self-healing process, PUT successfully self-healed the damaged area and recovered to its initial state within a short time. To quantitatively evaluate the self-healing efficiency of PU and PUT, a nano/micro scratching technique was used. As shown in Figure 2, PUT shows perfect self-healing performance (0.1-0 μm, 100%) in contrast to PU (0.11-0.4 μm, 64%), indicating BHMTHF PUD Polymers 2020, 12, x FOR PEER REVIEW 4 of 12

Sample
properties were analyzed using a differential scanning calorimeter (DSC), a thermal gravimetric analyzer (TGA), and a spectrophotometer (see Figure S2-S4 in the Supplementary Information for the DSC thermograms and TGA curves and the detailed optical properties of PU, PUB, PUT, and PUD). Comparison of PU to PUB, PUT, and PUD shows that all of the samples have similar transition temperatures and also show high transmittance and low haze and yellow index values, indicating that all of the functional diols used in this study did not significantly affect the initial thermal and optical properties of PU as listed in Table 2. It is generally known that a residual solvent in polymeric system can act as a plasticizer and significantly affect the final glass transition temperature. When the TGA results were analyzed to measure the amount of residual solvent, all of the TGA curves showed no observable change in weight, which indicates that all of the samples used in this study have no residual solvents.  Table 2. Optical properties (transmittance (Tr), yellow index (YI), and haze), thermal property (glass transition temperature (Tg)), and thicknesses of conventional polyurethane (PU), PU with aromatic moiety (PUB), PU with heterocyclic moiety (PUT), and PU with aliphatic moiety (PUD). All of the data points were obtained by averaging the results of three measurements. A self-healing experiment was conducted for PU, PUB, PUT, and PUD using the single scratch technique. Figure 1 shows the optical microscopy images before and after the self-healing process. It should be noted that PUB, PUT, and PUD contain an aromatic benzene moiety, heterocyclic tetrahydrofuran moiety, and aliphatic ethylene moiety, respectively, where all of these moieties have approximately the same molecular size. Interestingly, while PU, PUB, and PUD still retained traces of damage even after the self-healing process, PUT successfully self-healed the damaged area and recovered to its initial state within a short time. To quantitatively evaluate the self-healing efficiency of PU and PUT, a nano/micro scratching technique was used. As shown in Figure 2, PUT shows perfect self-healing performance (0.1-0 μm, 100%) in contrast to PU (0.11-0.4 μm, 64%), indicating DHE Table 2. Optical properties (transmittance (T r ), yellow index (YI), and haze), thermal property (glass transition temperature (T g )), and thicknesses of conventional polyurethane (PU), PU with aromatic moiety (PUB), PU with heterocyclic moiety (PUT), and PU with aliphatic moiety (PUD). All of the data points were obtained by averaging the results of three measurements. A self-healing experiment was conducted for PU, PUB, PUT, and PUD using the single scratch technique. Figure 1 shows the optical microscopy images before and after the self-healing process. It should be noted that PUB, PUT, and PUD contain an aromatic benzene moiety, heterocyclic tetrahydrofuran moiety, and aliphatic ethylene moiety, respectively, where all of these moieties have approximately the same molecular size. Interestingly, while PU, PUB, and PUD still retained traces of damage even after the self-healing process, PUT successfully self-healed the damaged area and recovered to its initial state within a short time. To quantitatively evaluate the self-healing efficiency of PU and PUT, a nano/micro scratching technique was used. As shown in Figure 2, PUT shows Polymers 2020, 12, 968 5 of 11 perfect self-healing performance (0.1-0 µm, 100%) in contrast to PU (0.11-0.4 µm, 64%), indicating that the heterocyclic group can provide an outstanding self-healing capability in conventional polyurethanes.

Sample
Polymers 2020, 12, x FOR PEER REVIEW 5 of 12 that the heterocyclic group can provide an outstanding self-healing capability in conventional polyurethanes.    Figure 1. Optical microscope images of PU, PUB, PUT, and PUD before and after the self-healing process. The loading force of 3 N was equally applied to all the samples and the self-healing temperature was set to be T g + 10 • C. The scale bar is 100 µm.
Polymers 2020, 12, x FOR PEER REVIEW 5 of 12 that the heterocyclic group can provide an outstanding self-healing capability in conventional polyurethanes.   The loading force of 50 mN was applied to both samples and the self-healing temperature and time were set to T g + 10 • C and 10 min, respectively. Figure 3 demonstrates an additional and crucial advantage of the introduction of the heterocyclic group to conventional polyurethanes. Various mechanical properties before and after the self-healing process were measured using a universal testing machine. PUT shows the best recovery of tensile strength, elongation at break, and Young's modulus compared to PU, PUB, and PUD. Considering the fact that all of the functional diol solutions have identical OH values and all of the functional diols have approximately the same molecular size, it is concluded that the heterocyclic group can also endow conventional polyurethane with superior recovery of mechanical properties.
Polymers 2020, 12, x FOR PEER REVIEW 6 of 12 Figure 3 demonstrates an additional and crucial advantage of the introduction of the heterocyclic group to conventional polyurethanes. Various mechanical properties before and after the self-healing process were measured using a universal testing machine. PUT shows the best recovery of tensile strength, elongation at break, and Young's modulus compared to PU, PUB, and PUD. Considering the fact that all of the functional diol solutions have identical OH values and all of the functional diols have approximately the same molecular size, it is concluded that the heterocyclic group can also endow conventional polyurethane with superior recovery of mechanical properties. To interpret these phenomena, a model system was used to propose the mechanism associated with improved self-healing performance and enhanced recovery of mechanical properties. Ethyl-N-methylcarbamate (U), tetrahydrofuran (T), and diethyl ether (E) were used to imitate PU, the heterocyclic group, and the heterolinear group, respectively. Figure 4 and Table 3 show the chemical structures of U, T, and E and the designations and compositions of the samples used for the model system. Based on previous reports [22], it is known that urethane forms three different types of hydrogen bonds (free (noN-Hydrogen bonded), single-hydrogen bonds, and double-hydrogen bonds) [21,23]. As shown in Figure 5a, the FT-IR spectrum of U100 (pure U) showed the double-hydrogen bonded C=O band at 1690.3 cm −1 , the single-hydrogen bonded C=O band (about 1715.0 cm −1 ) as a shoulder, and the free C=O band (about 1740.6 cm −1 ) as a weak shoulder (barely detectable). Interestingly, the peaks corresponding to the double-and single-hydrogen bonding both shifted to higher wavenumbers with increasing fraction of T, as shown in Figure 5a and Table 3. This can be explained as follows. As graphically shown in Figure 4a, when T approaches two U molecules that are connected by the weak hydrogen bonding between C=O and N-H, T will break the existing hydrogen bonds between the two U molecules due to the stronger specific interaction between the -O-in T and N-H in U. This will result in a stronger hydrogen bonding interaction between T and U and the force constant (K) of the dropped U will increase because the unpaired electrons in the C=O of the dropped U no longer interact with the partially positively charged hydrogen atom in N-H [20,[24][25][26][27][28][29][30]. As a result, based on the spring model for FT-IR spectroscopy, all of the peaks associated with the hydrogen bonded C=O shifted to higher wavenumbers as shown in Figure 5a.
Polymers 2020, 12, x FOR PEER REVIEW 7 of 12 To interpret these phenomena, a model system was used to propose the mechanism associated with improved self-healing performance and enhanced recovery of mechanical properties. Ethyl-Nmethylcarbamate (U), tetrahydrofuran (T), and diethyl ether (E) were used to imitate PU, the heterocyclic group, and the heterolinear group, respectively. Figure 4 and Table 3 show the chemical structures of U, T, and E and the designations and compositions of the samples used for the model system. Based on previous reports [22], it is known that urethane forms three different types of hydrogen bonds (free (noN-Hydrogen bonded), single-hydrogen bonds, and double-hydrogen bonds) [21,23]. As shown in Figure 5a, the FT-IR spectrum of U100 (pure U) showed the doublehydrogen bonded C=O band at 1690.3 cm −1 , the single-hydrogen bonded C=O band (about 1715.0 cm −1 ) as a shoulder, and the free C=O band (about 1740.6 cm −1 ) as a weak shoulder (barely detectable). Interestingly, the peaks corresponding to the double-and single-hydrogen bonding both shifted to higher wavenumbers with increasing fraction of T, as shown in Figure 5a and Table 3. This can be explained as follows. As graphically shown in Figure 4a, when T approaches two U molecules that are connected by the weak hydrogen bonding between C=O and N-H, T will break the existing hydrogen bonds between the two U molecules due to the stronger specific interaction between the -O-in T and N-H in U. This will result in a stronger hydrogen bonding interaction between T and U and the force constant (K) of the dropped U will increase because the unpaired electrons in the C=O of the dropped U no longer interact with the partially positively charged hydrogen atom in N-H [20,[24][25][26][27][28][29][30]. As a result, based on the spring model for FT-IR spectroscopy, all of the peaks associated with the hydrogen bonded C=O shifted to higher wavenumbers as shown in Figure 5a.   To analyze the FT-IR results quantitatively, the N-H bands were deconvoluted by fitting with mixed Lorentzian-Gaussian shaped components as shown in Figure 5b. The U100 spectrum initially contained the band of free N-H at 3440.4 cm −1 and the band of the N-H hydrogen bonded to U at 3350.0 cm −1 as listed in Table 3. When T was added to U, a new peak associated with the band of the N-H hydrogen bonded to T appeared due to the formation of stronger hydrogen bonding between U and T compared to that between two U molecules. As the weight fraction of T further increased, the peak position corresponding to the new peak shifted to lower wavenumbers. In addition, the To analyze the FT-IR results quantitatively, the N-H bands were deconvoluted by fitting with mixed Lorentzian-Gaussian shaped components as shown in Figure 5b. The U100 spectrum initially contained the band of free N-H at 3440.4 cm −1 and the band of the N-H hydrogen bonded to U at 3350.0 cm −1 as listed in Table 3. When T was added to U, a new peak associated with the band of the N-H hydrogen bonded to T appeared due to the formation of stronger hydrogen bonding between U and T compared to that between two U molecules. As the weight fraction of T further increased, the peak position corresponding to the new peak shifted to lower wavenumbers. In addition, the fitted peak area of the N-H hydrogen bonded with T increased and that of the N-H hydrogen bonded with U decreased simultaneously as listed in Table 3. All of these results clearly demonstrate that the heterocyclic group significantly enhances the hydrogen bonding interaction in conventional polyurethanes.
Lastly, the effect of the chemical structure of the functional diols containing heteroatoms on the resulting hydrogen bonding interaction was examined. Comparison of the FT-IR spectra of the blend of U and heterocyclic T to those of U and heterolinear E shows that the peak position of the double-hydrogen bonded C=O of U90T10 appeared at 1694.3 cm −1 , while that of U90E10 appeared at 1693.2 cm −1 as shown in Table 3 and Figure 5c. The difference between the corresponding peak positions of the UT blend and that of the UE blend increased as the amount of the heterofunctional diol increased as listed in Table 3. Considering the fact that both T and E have approximately the same molecular weight, it is concluded that the hydrogen bonding between U and T is much stronger than that between U and E. It is noted that crosslinked polyurethane is generally known to have structural irregularity. Interestingly, from the FT-IR results of PU, the areal ratio of hydrogen-bonded carbonyl groups to free carbonyl groups is 88.7 to 11.3, which implies that most urethane groups in crosslinked polyurethane are also connected by hydrogen bonding (see Figure S5 in the Supplementary Information for the N-H stretching region of PU). Based on all of the FT-IR results derived from the model system, it is strongly suggested that the heterocyclic group can significantly enhance the hydrogen bonding interaction in conventional polyurethane, resulting in a unique supramolecular network structure that can generate the main driving force to significantly improve both self-healing and mechanical properties as graphically shown in Figure 6. polyurethane are also connected by hydrogen bonding (see Figure S5 in the Supplementary Information for the N-H stretching region of PU). Based on all of the FT-IR results derived from the model system, it is strongly suggested that the heterocyclic group can significantly enhance the hydrogen bonding interaction in conventional polyurethane, resulting in a unique supramolecular network structure that can generate the main driving force to significantly improve both self-healing and mechanical properties as graphically shown in Figure 6.

Conclusions
In summary, a new self-healable polyurethane based on the heterocyclic group was prepared in this study. In contrast to conventional polyurethane, the heterocyclic self-healing polyurethane shows significantly enhanced self-healing performances. Furthermore, the heterocyclic self-healing polyurethane proposed in this study can overcome the traditional trade-off between self-healing and the deterioration of mechanical properties. This work can provide an avenue for the development of new functional materials for next-generation flexible displays.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: (a) Overall scheme for synthesis and 1 H NMR spectrum in CDCl3 and (b) GPC trace of polyol (poly(HEMA-r-BMA)), Figure  S2: DSC thermograms of PU, PUB, PUT, and PUD, Figure S3: Optical properties of PU, PUB, PUT, and PUD: (a) transmittance; (b) haze and yellow index. All data points were obtained by averaging three measurements, Figure S4: TGA curves of PU, PUB, PUT, and PUD, Figure S5: The N-H stretching region of PU in the FT-IR spectra. The spectrum was resolved into its components by the curve-fitting method.

Conclusions
In summary, a new self-healable polyurethane based on the heterocyclic group was prepared in this study. In contrast to conventional polyurethane, the heterocyclic self-healing polyurethane shows significantly enhanced self-healing performances. Furthermore, the heterocyclic self-healing polyurethane proposed in this study can overcome the traditional trade-off between self-healing and the deterioration of mechanical properties. This work can provide an avenue for the development of new functional materials for next-generation flexible displays.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4360/12/4/968/s1, Figure S1: (a) Overall scheme for synthesis and 1 H NMR spectrum in CDCl 3 and (b) GPC trace of polyol (poly(HEMA-r-BMA)), Figure S2: DSC thermograms of PU, PUB, PUT, and PUD, Figure S3: Optical properties of PU, PUB, PUT, and PUD: (a) transmittance; (b) haze and yellow index. All data points were obtained by averaging three measurements, Figure S4: TGA curves of PU, PUB, PUT, and PUD, Figure S5: The N-H stretching region of PU in the FT-IR spectra. The spectrum was resolved into its components by the curve-fitting method.